It is common for there to be a lot of “impulse” sessions, or sessions where the user is sitting at a bus stop, or waiting in an elevator, waiting for a friend to get back from the bathroom, etc and they just open and close an app.
These tend to jump as well when using more push notifications: https://www.quora.com/Is-there-a-way-to-calculate-median-session-duration-in-Google-Analytics
Sometimes, what Data Analyst will do, is dump the data into data warehouse and look at Median Session Length (but this requires a robust analytics team).
Otherwise, people will break it down by the duration blocks, where available (depending on the analytics tool).
For example, here is what a Duration block looks like in "Flurry":
Average Duration in Google Analytics:
There are some reliability issues with this particular metric in Google Analytics: https://www.visma.com/blog/analysis-reporting-average-session-duration-google-analytics/
In some cases customers will use two analytics tools so they can compare (eg: Mixpanel, ComScore, Flurry, etc...)